feat: Process spatial, temporal and genre (#938)

netwerk-digitaal-erfgoed · Jun 21, 2024 · 80a1340 · 80a1340
1 parent 45139cc
commit 80a1340
Show file tree

Hide file tree

Showing 5 changed files with 59 additions and 5 deletions.
diff --git a/shacl/register.ttl b/shacl/register.ttl
@@ -52,11 +52,15 @@ reg:DatasetShape
         reg:SchemaDateCreatedProperty,
         reg:SchemaDatePublishedProperty,
         reg:SchemaDateModifiedProperty,
-        reg:SchemaAlternateNameProperty ;
+        reg:SchemaAlternateNameProperty,
+        reg:SchemaKeywordsProperty,
+        reg:SchemaSpatialCoverageProperty,
+        reg:SchemaTemporalCoverageProperty,
+        reg:SchemaGenreProperty;
 
         # TODO
         # reg:SchemaVersionProperty, reg:SchemaManEntityOfPageProperty, reg:SchemaInLanguageProperty, reg:SchemaIsBasedOnUrlProperty, reg:SchemaCitationProperty,
-        # reg:SchemaGenreProperty, reg:SchemaKeywordsProperty, reg:SchemaSpatialCoverageProperty, reg:SchemaTemporalCoverageProperty, reg:SchemaIncludedInDataCatalogProperty,
+        # reg:SchemaIncludedInDataCatalogProperty,
 
 .
 
@@ -66,7 +70,7 @@ reg:DistributionShape
     sh:property
         # Required properties
         reg:DistributionContentUrlProperty,
-        reg:DistributionContentUrlIriProperty,        
+        reg:DistributionContentUrlIriProperty,
         reg:DistributionEncodingFormatProperty,
         # Recommended properties
         reg:SchemaDistributionNameProperty,
@@ -192,7 +196,7 @@ reg:SchemaDatasetProperty a sh:PropertyShape ;
     sh:path schema:dataset ;
     sh:class schema:Dataset ;
     sh:minCount 1 ;
-    sh:node schema:DatasetShape ;
+    sh:node reg:DatasetShape ;
     sh:name "Dataset(s) in de datacatalogus"@nl, "Dataset(s) in the data catalog"@en ;
     sh:message "Een datacatalogus dient minimaal een dataset te bevatten"@nl, "A data catalog should contain at least one dataset"@en ;
     sh:severity sh:Warning ;
@@ -334,6 +338,34 @@ reg:SchemaDistributionProperty a sh:PropertyShape ;
     # which is confusing to users.
 .
 
+reg:SchemaKeywordsProperty a sh:PropertyShape ;
+    sh:path schema:keywords ;
+    sh:minCount 1 ;
+    sh:severity sh:Info ;
+    sh:message "Een datasetbeschrijving kan trefwoorden bevatten"@nl, "A dataset description may contain keywords"@en ;
+.
+
+reg:SchemaSpatialCoverageProperty a sh:PropertyShape ;
+    sh:path schema:spatialCoverage ;
+    sh:minCount 1 ;
+    sh:severity sh:Info ;
+    sh:message "Een datasetbeschrijving kan een gebiedsaanduidingen bevatten"@nl, "A dataset description may contain spatial coverage"@en ;
+.
+
+reg:SchemaTemporalCoverageProperty a sh:PropertyShape ;
+    sh:path schema:temporalCoverage ;
+    sh:minCount 1 ;
+    sh:severity sh:Info ;
+    sh:message "Een datasetbeschrijving kan tijdsaanduidingen bevatten"@nl, "A dataset description may contain temporal coverage"@en ;
+.
+
+reg:SchemaGenreProperty a sh:PropertyShape ;
+    sh:path schema:genre ;
+    sh:minCount 1 ;
+    sh:severity sh:Info ;
+    sh:message "Een datasetbeschrijving kan genres bevatten"@nl, "A dataset description may contain genres"@en ;
+.
+
 ##
 ## DCAT
 ##

diff --git a/src/query.ts b/src/query.ts
@@ -17,6 +17,9 @@ const dateModified = 'dateModified';
 const language = 'language';
 const source = 'source';
 const keyword = 'keyword';
+const spatialCoverage = 'spatialCoverage';
+const temporalCoverage = 'temporalCoverage';
+const genre = 'genre';
 const mainEntityOfPage = 'mainEntityOfPage';
 const version = 'version';
 const includedInDataCatalog = 'includedInDataCatalog';
@@ -74,6 +77,9 @@ export const constructQuery = `
       dct:source ?${source} ;
       dcat:keyword ?${keyword} ;
       dcat:landingPage ?${mainEntityOfPage} ;
+      dct:spatial ?${spatialCoverage} ;
+      dct:temporal ?${temporalCoverage} ;
+      dct:type ?${genre} ;
       owl:versionInfo ?${version} ;
       dct:isPartOf ?${includedInDataCatalog} ;
       dct:publisher ?${publisher} ;
@@ -151,6 +157,9 @@ export const constructQuery = `
         OPTIONAL { ?${dataset} dct:language ?${language} }
         OPTIONAL { ?${dataset} dct:source ?${source} }
         OPTIONAL { ?${dataset} dcat:keyword ?${keyword} }
+        OPTIONAL { ?${dataset} dct:spatial ?${spatialCoverage} }
+        OPTIONAL { ?${dataset} dct:temporal ?${temporalCoverage} }
+        OPTIONAL { ?${dataset} dct:genre ?${genre} }
         OPTIONAL { ?${dataset} owl:versionInfo ?${version} }
         OPTIONAL { ?${dataset} dct:isPartOf ?${includedInDataCatalog} }
         OPTIONAL { ?${dataset} dcat:landingPage ?${mainEntityOfPage} }
@@ -242,6 +251,9 @@ function schemaOrgQuery(prefix: string): string {
     OPTIONAL { ?${dataset} ${prefix}:isBasedOn ?${source} }
     OPTIONAL { ?${dataset} ${prefix}:isBasedOnUrl ?${source} } 
     OPTIONAL { ?${dataset} ${prefix}:keywords ?${keyword} }
+    OPTIONAL { ?${dataset} ${prefix}:spatialCoverage ?${spatialCoverage} }
+    OPTIONAL { ?${dataset} ${prefix}:temporalCoverage ?${temporalCoverage} }
+    OPTIONAL { ?${dataset} ${prefix}:genre ?${genre} }
     OPTIONAL { ?${dataset} ${prefix}:version ?${version} }
     OPTIONAL { ?${dataset} ${prefix}:includedInDataCatalog ?${includedInDataCatalog} }
     OPTIONAL { ?${dataset} ${prefix}:mainEntityOfPage ?${mainEntityOfPage} }

diff --git a/test/datasets/dataset-dcat-valid.jsonld b/test/datasets/dataset-dcat-valid.jsonld
@@ -58,6 +58,11 @@
     }
   ],
   "dct:language": ["nl-NL", "en-GB"],
+  "dct:spatial": {
+    "@id": "https://sws.geonames.org/2750405/"
+  },
+  "dct:temporal": "1939/1945",
+  "dct:type": "Interesting data",
   "dcat:distribution": [
     {
       "@type": "dcat:Distribution",

diff --git a/test/datasets/dataset-schema-org-valid.jsonld b/test/datasets/dataset-schema-org-valid.jsonld
@@ -17,10 +17,15 @@
     "@language": "en"
   },
   "keywords": "alba amicorum",
+  "genre": "Interesting data",
   "url": "https://www.kb.nl/bronnen-zoekwijzers/kb-collecties/moderne-handschriften-vanaf-ca-1550/alba-amicorum",
   "identifier": "http://data.bibliotheken.nl/id/dataset/rise-alba",
   "license": "http://creativecommons.org/publicdomain/zero/1.0/",
   "inLanguage": ["nl-NL", "en-GB"],
+  "spatialCoverage": {
+    "@id": "https://sws.geonames.org/2750405/"
+  },
+  "temporalCoverage": "1939/1945",
   "publisher": {
     "@type": "Organization",
     "@id": "https://example.com/publisher",

diff --git a/test/fetch.test.ts b/test/fetch.test.ts
@@ -77,11 +77,11 @@ describe('Fetch', () => {
 
     expect(datasets).toHaveLength(1);
     const dataset = datasets[0];
-    expect(dataset.size).toBe(31);
 
     const dcatEquivalent = await dereference(
       'test/datasets/dataset-dcat-valid.jsonld'
     );
+    expect(dataset.size).toStrictEqual(dcatEquivalent.size);
     expect(dataset.toCanonical()).toStrictEqual(dcatEquivalent.toCanonical());
 
     expect(